Preliminary results from quite a few samples, which are too boring and too numerous to post:
Phamiya 645D with 150 F3.5 MF lens shot tethered and focussed using Live View.
NO MUP (the reason for this is that I want to see how this works for portraits at the moment, rather than my mainstream landscape stuff so I will test with MUP later). This also explains why all the tests were done with vertical orientation, which is more challenging because of cantilever effects.
Shutter speeds of 1/200th through 1/12th second at a fixed ISO of 50.
Target was a lens test chart on a wall about three metres away.
Tripod was a Gitzo GT 3541LS and a cable release was used. A hard laid tiled floor was used. There was no breeze.
Head combo 1: Cube with Arca L plate (kindly supplied by Martin Vogt of Arca) in VERTICAL position on a flat cube.
1/200th Good
1/100th Good
1/50th Not great
1/25th No way
1/12th Not great but the best of the three tested at this shutter speed and I might risk it at a pinch.
Head combo 2: Cube with Arca L plate in HORIZONTAL position but with the Cube tilted 90 degrees
1/200th Good
1/100th Good
1/50th Good
1/25th 90% good
1/12th Wouldn't use it
Head combo 3: Manfrotto 410 with standard plate and oriented at 90 degree tilt so as to give vertical FOV
1/200th Good
1/100th Good
1/50th Good
1/25th 80% good
1/12th Wouldn't use it.
'Wouldn't use it' means IMHO that there's a two pixel blur or worse and that the shot is clearly less sharp at 50% on screen than the reference shots taken at 1/200th.
As for the possibility of focus shift being the cause of apparent blur (rather than shake): I haven't tested this lens for it because I've never suspected it, nor was there any pattern in these tests to suggest it.
I will test this rig further especially in the horizontal orientation and also with MUP and various delays between MUP and shutter release. In the meantime the Cube bests the 410, just, when used with the L plate in horizontal position but the cube itself providing the 90 degree tilt. If the vertical position option of the Arca plate is used, there seems to be less stability even at 1/50th second but surprisingly at 1/12th second this is the best permutation - but still not good enough really.
So the cube wins, just and by a tiny margin, but I don't trust the L plate in vertical position: I prfer it horizontally deployed with the cube itself providing 90 degrees of tilt.
All the usual health warnings apply but I did try all this a few times and am certainly confident enough with the results to act on them myself!
Hope that helps, or even mildly interests, someone! I think that the Arca L plate has some issues in vertical orientation, because used in that way the camera is quite heavily cantilevered way from the point of attachment to the plate. Admittedly the cube itself, when tilted to 90 degrees, has a lot of cantilever stress but it seems to handle it well. The fact that the best results at 1/12th are form the Cube with L plate in vertical position is most likely due to some aspect of harmonic vibrations or sympathetic frequencies in the rig. The Manfrotto does extremely well when the price comparison is made, falling only very slightly behind the Cube's best showing and even then only at the 1/25th shutter speed BUT it is less elegant, annoying and clunky to use (though actually it can be faster when you get used to it) and it is both heavier and more angular to pack.
Best
Tim